Wi!Mi Expert System Shell as the Novel Tool for Building Knowledge-Based Systems with Linear Computational Complexity

摘要

Expert systems (ES) are very effective tools for solving various complex problems, which usually require human intelligence. In fact, these systems emulate a human expert's decision-making process. However, development of various ESs from scratch is a tedious and costly process. This article describes a novel ES building tool This tool belongs to the class of the so-called Wi!Mi ES shells-programs that allow significantly simplify and accelerate the ES development process. It uses a new knowledge representation model, which is based on the bipartite graphs as well as on the significantly improved but simple inference engine. This tool makes it possible to process large systems involving millions of variables during several seconds. This article presents both theoretical basis and inner workings of the Wi!Mi tool, as well as several examples of various subject domains. Finally, this tool is evaluated by testing on the prototype including up to 3 million variables and rules.